Conclusion

In conclusion, the thickness-rule test allowed us to evaluate the performance of our CNC machine using a 0.1 mm V-bit with a 30° angle. Although it was possible to cut very thin lines, poor clamping and the absence of dogbones caused the board to move during machining, resulting in an imprecise cut. To improve future results, it is crucial to secure the material firmly and include dogbones to prevent unwanted movement during the cutting process. Unfortunately, the TT3018 machine had connection errors that made further testing impossible, but with better clamping and dogbones, more accurate results are expected in future tests.

Individual assignment:

• Make and test a microcontroller development board that you designed.

FLATCAM PCB PRODUCTION

This time I used FlatCAM. I loaded the GERBER file of my test board, configured the tools for engraving and cutting, and finally exported the machining files to send them to the CNC. Z-axis calibration was done with both automatic and manual methods to compare results. Unfortunately, the board moved during machining, which resulted in an imprecise cut, but with better clamping and the inclusion of dogbones, more precise results are expected in future tests.

We opened the GERBER file previously saved from KiCad and positioned it at positive origin coordinates.

We create the toolpath by setting the dimensions of our bits and the pass overlap percentage.

We configure the highlighted parameters with the corresponding bits, feed rates, and RPM of the CNC we are using.

We use the 2mm carbide PCB cutting milling cutter and the 0.1mm spear point milling cutter for engraving.

We export the G-code generated by the design software to the CNC.

THE TT3018 CNC IS NOT RECOMMENDED!!

Although there are several versions and quality levels, the one I used had connection errors and electrical noise, causing the machine to freeze during machining. I also read that the same thing happens to other people. I still tried and used Candle to load the G-code.
What impressed me was its height-map calibration for PCB boards, performing Z probing across the entire work area.

We open our G-code in Candle and move to the work area. We install the automatic Z probe as shown in the photo. We also use a 0.1 mm, 30° V-bit for track engraving and a 2 mm flat end mill for cutting. Then we click Create to generate the Z height map. A window opens to configure the height map; in this case I set it to 5 mm depth and 5 mm step. Next, another window opens to configure the area to map; I used a 10 mm margin on each side. Finally, a new window appears to configure the map type; I left it as a normal height map. The height map is then displayed and can be saved for future machining jobs.

We check the height map option and then click "start" to begin machining.

THE TT3018 CNC IS NOT RECOMMENDED!! I MADE MANY ATTEMPTS, BUT THE MACHINE KEPT STOPPING AND DISCONNECTING. I TESTED MANY CODES WITH SHORTER TOOLPATHS, BUT NOTHING WORKED. I STILL NEED TO COMPLETE THE ASSIGNMENT, SO I WILL TRY ANOTHER MACHINE.

I CAME TO FABLAB U.P TO USE THEIR SNAPMAKER CNC

In record time I had to learn how to use the SNAPMAKER: place my PCB, secure it, change tools, and level the surface.

I brought my G-code and loaded it into the SNAPMAKER.

The results are good, but I think I needed more engraving spacing. The process took too long, so I think I will do it with laser engraving instead.... XDDDD

SOLDERING IS NOT MY THING!! XD

The socket for the XIAO was easy for me to solder. Everything was fine up to that point... but the ending was F.

WOW, soldering on a full copper board is complicated. I think it is necessary to remove everything that is not part of the traces.

I BETTER MAKE MY PCB WITH LASER ENGRAVING

This is the final schematic with which I will make my laser PCB, a button and LED were added in relation to week 6 of "Electronics Design"

This is how my PCB design would look, with the traces and cuts for the laser. The position of the external power connectors was also changed.

With FlatCAM already open from the previous step, we now export our design in SVG format.

We import the SVG of our board into the EZCAD2 software to engrave it with the laser. We set the hatch to cross pattern and use the following parameters: 500 mm/s speed, 70% power, and 45 kHz frequency in 2 passes.

Wowww...que facil se me hace soldar asi!!! XD

The PCB was tested with the XIAO RP2040, turning on the LED with the button. Below is the code and video.

Code: PCB Test (Button + LED)   Original code: Gemini

// Pin definitions
const int pinLED = D2;    // Pin where the LED is connected
const int pinBoton = D3; // Pin where the button is connected

void setup() {
  // Configure the LED pin as output
  pinMode(pinLED, OUTPUT);
  
  // Configure the button pin as input with internal PULLUP resistor
  // This prevents the pin from "floating" and detects a 0 (LOW) when pressed
  pinMode(pinBoton, INPUT_PULLUP);
}

void loop() {
  // Read the button state
  int estadoBoton = digitalRead(pinBoton);

  // Since we use INPUT_PULLUP, the state will be LOW when pressed
  if (estadoBoton == LOW) {
    digitalWrite(pinLED, HIGH); // Turn the LED ON
  } else {
    digitalWrite(pinLED, LOW);  // Turn the LED OFF
  }
}
                                                                

Conclusions:

• I learned to use FlatCAM to prepare GERBER files and generate engraving and cutting G-code for PCB production, correctly configuring V-bits and depth/angle parameters.
• The TT3018 CNC turned out to be unreliable for PCB production: it has connection errors, electrical noise, and freezes during machining, which made it impossible to complete the work on that machine.
• I learned the importance of proper material clamping and using dogbones; without them, the board moves during machining, resulting in inaccurate traces and failed cuts.
• The Snapmaker proved to be a more reliable alternative: I learned to clamp the board, change tools, level the surface, and load G-code in record time to finish the assignment.
• I learned that soldering SMD components on a full copper board is very difficult; it is necessary to remove excess copper that is not part of the traces to avoid short circuits and make soldering easier.
• Fine-pitch soldering is still a personal challenge; the XIAO socket was manageable, but smaller components require more practice and better tools.

FILES

PCB HEXAMODULAR CNC Download .CNC

HEXAMODULAR SVG Download .SVG